The Design of Lighting

(1)

(2)

(3)

(4)

The design of lighting

Peter Tregenza and David Loe

E & FN SPON

An Imprint of Routledge

(5)

This edition published in the Taylor & Francis e-Library, 2009. To purchase your own copy of this or any of

Taylor & Francis or Routledge’s collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk.

Simultaneously published in the USA and Canada by Routledge

All rights reserved. No part of this book may be reprinted or reproduced or utilized in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing

from the publisher.

British Library Cataloguing in Publication Data

A catalogue record for this book is available from the British Library ISBN 0-203-22363-2 Master e-book ISBN

ISBN 0-203-27781-3 (Adobe ebook Reader Format) ISBN 0 419 20440 7 (Print Edition)

(6)

List of tables

1.1 The units of lighting

4.1 Typical lamp characteristics

4.2 Short checklist for choosing a lamp

4.3 Short checklist for choosing a luminaire

6.1 Checking an interior lighting design

7.1 A ‘normal’ daylit room

8.1 Some factors affecting colour choice

8.2 Some strategies for coordinating colours

9.1 Typical recommended task illuminances

9.2 Four practical checks for glare at a workplace

9.3 Briefing checklist for task lighting

10.1 Checklist for views

10.2 Rectangular rooms with view windows in one wall: minimum glazed areas

10.3 Room appearance and average daylight factor: values associated with rooms in temperate climates

11.1 Illuminance ratios for displaying objects

11.2 Basic contrasts in display

11.3 Some display lighting techniques

11.4 Typical maximum illuminances and light exposures recommended for long-term conservation

(11)

12.1 Functional checklist for exterior lighting

12.2 Typical recommended exterior night-time illuminances

12.3 Briefing checklist for floodlighting a building

14.1 Luminaire maintenance factors (LMF) for a cleaning interval of one year

14.2 Room surface maintenance factors (RSMF) for a cleaning interval of one year

14.3 Typical maintenance factors adopted in daylighting calculations

14.4 Checklist of other environmental factors

16.1 _{Plan distance, d, between window reference and skyline points; azimuth,}

, and elevation, , of skyline points

(12)

Preface

This book is an introduction to lighting in buildings, written for architects, interior designers and building services engineers. It is planned to be a reference for practitioners and a textbook for students. It covers daylighting and electric lighting, and introduces the use of colour.

There are three parts. The first, The Technical Background , gives the framework of physical and human factors used by the lighting designer: the ways in which light and colour are described and quantified, the fundamentals of vision, and the sources of light. These are set out in short chapters, which can be either read as an introduction, skipped by the knowledgeable, or used for reference. The aim is to give a concise description of topics that may make up the lighting syllabus of a degree in architecture, and to include some of the elements of interior design and environmental engineering courses.

The second part, Designing , is the main section of the book. It describes the needs and preferences of people in buildings, and shows that there are many criteria of good lighting, ranging from the perceived character and architectural form of a space to the detailed requirements of task performance and comfort. Specific chapters cover design of the overall light and colour of a room, windows, lighting for work and lighting for display. Exterior lighting of buildings is described, and further chapters deal with emergency lighting, energy use and other design constraints.

The third part, Calculations , sets out procedures required in practice. Taking some of the most frequently used daylighting and electric lighting calculations, it gives for each a step-by-step synopsis of the method and a realistic worked example. The final chapters list typical data and sources of further information.

A feature of the text is a sequence of tables based on current design codes and data, summarizing good practice and providing checklists. The Design of Lighting is planned to be used in conjunction with national standards and codes of practice, giving the designer some of the ideas that lie behind current recommendations, particularly where these have developed out of research in the last ten years.

But underlying all is the belief that lighting is a visual art, no more to be understood from textbooks or accomplished by calculation than any other aspect of architectural de-sign. Learning is primarily by doing—by designing buildings in brightness and colour, by observing and recording. The aim of the whole book is to provide a framework for this.

(13)

The authors gratefully acknowledge the following sources:

Concord Sylvania, Figures 7.9, 7.10, 8.2, 11.3, 11.4, 12.1, 12.2, 12.6 Judith Torrington, Figure 9.7

Peter Blundell Jones, Figure 7.2

Philips Lighting, Figures 7.5, 12.4, cover photograph Peter Lathey, Figures 9.3, 9.5

The National Gallery, London, Figures 6.1, 7.3, 7.8, 7.13

Much of the information given in the book is based on printed sources. In particular, the guideline tables of criteria and design recommendations are largely derived from data in publications from the Commission Internationale de l’Eclairage and the Chartered Institution of Building Services Engineers. Several figures, particularly in Part One, draw heavily on widely reproduced illustrations in standard texts on vision and on lighting technology.

The authors also acknowledge help and advice from colleagues at the University of Sheffield, at University College London, and at the Building Research Establishment. They are grateful to their families and to friends in research and in the lighting industry throughout the world for help and encouragement.

(14)

(15)

The Technical Background

This is an outline of the physical ideas underlying lighting design. It shows how light and colour can be described, how the human eye responds, how light is produced electrically, and how the light from the sky can be predicted.

It is set out in short chapters, which can be read as an introduction or used for reference.

(16)

(17)

Describing light

Light is a flow of energy. Like radiant heat, radio waves and X-rays, it is part of the electromagnetic spectrum, and can be described in terms of wavelength and power. But what we see as light can be mixed from many colours, and there is no one-to-one link be-tween the spectral distribution of radiation and human perception of brightness and hue.

For this reason, light is defined uniquely by the response of the human eye. It has its own set of units, which allow it to be quantified, and which are linked to other units of power (such as watts) only by a standardized mathematical description of visual sensitivity. There are four interrelated units. These describe the flow of light, its intensity in space, illuminance at a point, and the luminance of a surface.

FLOW AND INTENSITY

The first unit of light is the lumen (lm). It describes luminous flux, the total flow of light from a source, just as the flow of heat from a radiator can be described in watts (W). The output from a lamp is given in lumens. For example:

100 W incandescent lamp 1360 lm

58 W fluorescent tube 5200 lm

400 W high-pressure sodium lamp 48 000 lm

These are typical values, but light output depends on the details of a lamp’s construction, and it decreases as the lamp ages.

The relationship between a lamp’s light output and its electrical input is known as

luminous efficacy, measured in lumens per watt. This depends on the physical efficiency

of the lamp and the spectral distribution of its output. An incandescent lamp has a low efficacy because most of its power is radiated as heat, in the infrared part of the spectrum, not as light. Efficacies of the three examples are

100 W incandescent lamp 13.6 lm/W

58 W fluorescent tube 90 lm/W

400 W high-pressure sodium lamp 120 lm/W

The performance of a luminaire (a light fitting) depends not only on the total amount of light emitted but also on how this is distributed. It could be concentrated in a narrow beam or diffused broadly. The term luminous intensity (or just intensity) is used to describe the flow of light in a given direction. Intensity is measured in candelas (cd). It is calculated from the number of lumens divided by the angular size of the beam, measured in steradians: a candela is 1 lumen per steradian. A spotlight used in the home to il-luminate a picture might have a luminous intensity of 3000 cd along the axis of its beam.

(18)

4 The design of lighting

A steradian describes a solid angle, the spatial equivalent of the radian. One steradian is the solid angle at which the area on the surface of a sphere is equal to the radius squared (as shown in Figure 1.1). It is a ratio, and its use in the definition of the candela simplifies calculations.

Figure 1.1 Solid angle . Its size is A/d2 steradians.

ILLUMINANCE AND LUMINANCE

Illuminance, measured in lux (lx), is the amount of light falling on a

surface (luminous flux density). One lux is given by one lumen falling evenly on a square metre. Here are some typical values of illuminance:

from a candle 1 m away 1 lx

on desks in a general office 500 lx

on the ground from an overcast sky 10 000 lx from the sun and bright sky in summer 100 000 lx

Standards for lighting are most frequently given as the required lux on the working plane, where the working plane is taken to be a horizontal surface at task level across the room.

The apparent brightness of a surface depends partly on the adaptation state of the eye (which is described in Chapter 3), and partly on the quantity of light reaching the eye from the surface. The term luminance (objective brightness) is used to define the physical quantity. The magnitude of the luminance depends on two things: the intensity of light from the surface in the direction of the viewer, and the projected area of the surface emitting or reflecting this light. The smaller the surface area, the brighter it must be to produce a given intensity.

The unit of luminance is therefore the candela per square metre. Typical values are:

white paper on an office desk _{130 cd/m}2

overcast sky _{3000 cd/m}2

white paper in strong sunlight _{25 000 cd/m}2

Sky luminance is usually given in cd/m2 even though a square metre of sky cannot be visualized.

(19)

ALL THE UNITS ARE RELATED

The units are linked together by measures of solid angle and area, as shown in Table 1.1.

Table 1.1 The units of lighting

Unit

Luminous flux F lumen (lm)

Luminous intensity: flux/solid angle l candela (cd)

Illuminance: flux/area E lux (lx)

Luminance: intensity/projected area L _{candela per square metre (cd/m}2

)

Two basic equations derived from this interrelationship form the foundation for nearly all lighting calculations. The first, which is illustrated in Figure 1.2, gives illuminance in terms of luminous flux and receiving surface area:

A surface in a room receives a fraction, p, of all the light emitted from the lamps in the room. The flux from the lamps is F lumens; the working plane area is A square metres. Then the average illuminance in lux on the surface is

(1.1)

This is the basis of the lumen method, which is described in detail in Chapter 16, Example (d). In practice the fraction p is usually taken to have two components: a

utilization factor, which takes into account the directionality and reflection of light

within the space; and a maintenance factor, which indicates reduction of illumination during the lifetime of the lighting installation.

If a source of light is very small, effectively a point in space, the illuminance that it produces on a surface depends on intensity, distance and angle of incidence. This is the second basic equation:

Symbol Quantity

The intensity of the light in a particular direction from the source is I candelas; the light travels distance d metres and falls on a surface at an angle (this angle of incidence is measured between the direction of light and the normal, or

perpendicular, to the surface). The illuminance on the surface is, in lux,

(20)

(21)

This equation embodies the inverse square law and Lambert’s law (which says that the amount of light falling on a surface is proportional to the cosine of the angle of incidence). See Figure 1.3.

Figure 1.3 Illuminance from a point source.

No real lamp is a point source, but the equation can be used with negligible error when the dimensions of the source are small in relation to the distance d (a maximum source size of d/5 is often taken to be the limit) and when the rays are not focused by an optical system. Chapter 16 shows how the equation is used to calculate illuminance from sources such as spotlights.

REFLECTION AND TRANSMISSION

The fraction of the incident light that is reflected back by a surface is the reflectance, denoted by the Greek letter rho ( ). Reflectance is a value between zero and one: =0 if the surface is perfect black and therefore absorbs all light; =1 if all incident light is

reflected.

Figure 1.4 Specular reflection.

The direction of the reflected beam depends on the nature of the surface. A specular

reflector is a mirror: a ray is reflected in the same plane as the incident ray and at an

angle equal to the angle of incidence, as in Figure 1.4. The luminance of a perfectly specular surface does not depend on its illuminance but on the luminance of the objects

(22)

seen reflected. A diffuse reflector is a matt surface that scatters light evenly in all directions. The luminance of a perfect diffuser is proportional to its illuminance, and is the same from every angle of view.

No real surface is either entirely specular or diffusing. Sometimes it is useful to give two reflectances for a material, specular and diffuse; and for scientific study a directional reflectance function may be necessary, specifying how a beam of light at any incident angle is scattered. For lighting calculations in buildings it is usually enough to assume that diffuse reflection is dominant and that a single value of for a particular material describes its total reflectance. Typical reflectances are:

white paper _0.8

clean concrete _0.4

dark wood _0.1

With a diffuse reflector, as in Figure 1.5, luminance (cd/m2) is directly proportional to illuminance and reflectance:

(1.3)

Figure 1.5 Luminance of a diffuse reflector.

The link between luminance, area and intensity, as in Figure 1.6, completes the chain:

If the area of a patch of surface is A m2, its luminance is L cd/m2, and it is facing an angle ø away from the line of sight, then the intensity towards the viewer is

(23)

Figure 1.6 Intensity and luminance.

For this intensity to be used to calculate the illuminance on another surface with equation (1.2), the size of the bright patch must be small in relation to the distance. This may mean that the patch must be subdivided into small zones and the calculation repeated for each of them. Alternatively, for advanced calculations there are formulae that give directly the illuminance from line and area sources.

Transmittance is the fraction of light that passes through a material. It also is a

number between zero and one, and is denoted by the Greek letter tau ( ). Sometimes it is useful to specify diffuse transmittance, the fraction of a beam that is uniformly scattered, and regular transmittance, the fraction that remains as a geometrical ray.

In all real transparent materials there is some reflection at surfaces, some absorption and some scattering. The sum of the total reflectance and the total transmittance must be a number between zero and one. With a material such as glass, the fractions that are reflect-ed and transmittreflect-ed depend on the angle of incidence. When a beam strikes a glass surface at a glancing angle it is mainly reflected; when it is perpendicular to the surface most of it passes through. For simple calculations of window performance an average transmittance is used, a weighted mean over all directions of incidence. Typical values of daylight transmittance are:

clear 6 mm sheet glass without dirt _0.8

clear double glazing with average dirt _0.5

The transmittance of glass to light differs from its transmittance of total solar radiation or of radiant heat because the transparency of the material varies with wavelength. Separate values are used for daylighting and solar gain calculations. But all absorbed energy, whatever the wavelength of the incident radiation, increases the temperature of a material. All the light that enters a room is ultimately absorbed by surfaces and is therefore a thermal gain.

Chapter 16 gives typical reflectances of building materials and some examples of glazing transmittance.

(24)

2 Describing colour

It is not enough for a designer to describe a colour just as ‘red’, or even ‘strong yellowish red’: colour names have different meanings to different people. To match colours accurately or to achieve subtle mixtures requires a far more precise method of specification.

But practicality is not the only reason for developing a conceptual structure of colour. Having a framework for visualizing chromatic relationships can be creatively helpful: for instance, the sequence given in Chapter 8 for selecting colours is based on a concept of colour classification.

SURFACE COLOURS

The simplest method of ensuring standardization is to have a reference set of actual colour samples. A manufacturer’s card of paint colours or a swatch of fabric samples lets the consumer make an unambiguous choice: all that is necessary is to quote the number or name by which the item is labelled.

Such a set of samples is called a colour atlas. Its use is not, however, foolproof. In any colour manufacturing or printing process there is variation from one batch to another, so different copies of a reference set are never exactly identical. But, more importantly, the appearance of pigments depends on the nature of the light falling on them. Two surface colours can appear similar under one illuminant and significantly different under another (a phenomenon known as metamerism). A colour atlas must always be viewed under the same type of illumination, and this is normally assumed to be light of a continuous spectrum such as daylight.

But an atlas consisting of only a set of samples cannot itself provide exact specification of intermediate colours. The solution is to create axes along which individual points are set, just as longitude and latitude can be used to specify a place on a globe. Colour, though, requires three dimensions.

These three dimensions are most commonly called hue, value and chroma. Hue is a point on the colour circle, value varies with the reflectance of the coloured surface, and chroma varies with the saturation of the colour. The dimensions can be visualized when a strong pigment is added to a neutral grey paint. The hue is given by the pigment, the chroma by the amount of pigment added, and the value by the neutral paint that would have the same reflectance as the final mixture.

(25)

Figure 2.1 The Munsell hue circle.

The terms hue, value and chroma are those of the Munsell system, first published in 1905 by A.H.Munsell, a Boston art teacher. In this system, the hue circle is divided into ten main segments: red, yellow-red, round to purple and red-purple, as in Figure 2.1. Each segment is further divided into ten divisions around the pure hue. The hue is then coded: 5Y, for example, is pure yellow.

The hue circle can be made into a colour disc by placing a grey point in the centre and increas-ing saturation radially from this neutral to the pure hue. The distance from the centre is Mun-sell’s indicator of chroma: zero indicates neutral, a high number an intensely saturated colour.

Value is the third dimension. Black has a value equal to zero, white a value of 10. The re-flectance of a coloured surface is given approximately by a simple function of the Munsell value:

(2.1)

(26)

Figure 2.2 The Munsell colour solid, cut to show scales of value and

chroma across a hue.

Each point within the Munsell solid is a unique colour and has a complete reference in the form

hue value/chroma

except for neutrals, which are denoted by N, with only value. For example:

light yellow 5Y 8/0.5

deep red-purple 7.5 RP 2/4

dark grey N 3

Within each dimension equal increments are intended to indicate equal steps in perceived contrast, although the spacing varies between dimensions. The solid is asymmetrical because the values of apparently fully saturated hues differ: a strong yellow paint looks lighter and has a higher reflectance than a saturated blue.

The colour solid can be imagined as a fruit, with the core white at the top and black at the base, and the skin graduated to give a surface of spectral colours. Unlike an apple, the inside also is coloured and graduated: a worm burrowing across in a straight line might begin eating yellow, pass through mid-grey, and emerge from dark blue.

(27)

Figure 2.3 The NCS hue circle.

The Natural Colour System is an alternative method of defining surface colours. In this system, a colour is defined by its position within a set of six elementary colours: white and black, plus the chromatic elements yellow, red, blue and green. The chromatic elements are set at the quarter points of a colour circle, and an intermediate hue is given as a percentage of the distance along the arc between each adjacent pair. An orange, halfway between yellow and red, is denoted Y50R, while a red with only a tinge of yellow is Y90R. The circle of chromatic colours is shown in Figure 2.3.

A colour that is not fully saturated is denoted by its position in a triangle formed between white, black and a chromatic colour. The edges of the triangle form the axes of whiteness, blackness and chromaticness. The colour solid is thus a double cone, as in Fig-ure 2.4. The top point is white, the base point is black and the colour circle forms the rim.

Figure 2.4 The NCS colour solid, cut to show scales of blackness and

chromaticness.

The sum of blackness, whiteness and chromaticness at any point equals 100%. In notation blackness and chromaticness are given two digits each while whiteness is omitted because it is a redundant dimension. The format used is

(28)

blackness chromaticness—hue

Deep red-purple could be described by 7315—R24B. That is, 73% blackness, 15% chromaticness, with a hue 24% of the distance along the arc between red and blue in the colour circle.

The NCS equivalents of the other Munsell examples are

dark grey 7501—R97B

light yellow 2002—Y03R

Unlike the Munsell system there is no simple relationship between equal increments along the dimensions and equal steps in perceived colour difference, nor is there an approximate link to reflectance.

Within the UK a system for specifying surface colour is given in British Standard BS 5252. This is an atlas of 237 colours on a reference system that uses the three dimensions of hue, greyness and weight. The first of these is comparable with Munsell hue, while greyness is comparable with chroma except in the designation of white and black. Weight, though, is based on subjective lightness instead of reflectance, and therefore does not correlate well with value because colours of the same reflectance in different saturated hues do not appear equally light.

Hues, in the BS notation, are designated by a two-digit even number: 00 represents neutral, 02 red-purple, 04 red, and then in steps round the colour circles to 24 purple. Greyness is given by a letter from A (grey) to E (clear or pure hue). Weight is given by a two-digit odd number, 01 indicating a very low weight.

The system was intended to help designers to select colour combinations, but it has not received widespread support.

The format of BS 5252 notation is

hue greyness weight

and the examples become

dark grey 00 A 13

light yellow 10 A 03

(29)

Figure 2.5 Matching lights: additive colour.

If, as in Figure 2.5, three different coloured lamps are focused to shine together onto a white surface, the resulting mixture of light can be changed in colour by altering the relative outputs of the sources. It might be possible to adjust the combination so that the patch of light from a fourth, different, lamp appears identical in colour.

To be able to match a broad range of other colours, the source lamps, the primaries, must be strongly saturated and be widely spaced in hue around the colour circle (red, green and blue, for example). Then, if the primaries are standardized, another colour could be defined by the combination that gives the same appearance. It could be written

If these are expressed in relative quantities so that , the mixture can be plotted as a point in a triangle, as Figure 2.6.

The superimposition of lights gives additive colour mixing: an apparently white light can be obtained by combining the primaries. (Subtractive colour mixing occurs when paints are mixed: a combination of red, blue and green pigments approaches black because each colour absorbs part of the total incident light. Subtractive mixing occurs also when light-transmitting materials such as filters are combined.)

But not every colour can be matched by a set of three primaries. Those excluded are other saturated spectral sources. A combination of light from red and green lamps produces yellow, but a more intense colour can be obtained from a pure yellow source. A matching combination could occur only if the pure yellow were desaturated with some light from the blue primary—so in the rgb specification the blue component would take a negative value.

(30)

Figure 2.6 The CIE x,y chromaticity diagram. The triangle formed by

plotting the coordinates of three actual lamp colours shows the range of colours that can be matched by mixing these.

The CIE chromaticity diagram (Figure 2.6) is based on three imaginary sources with standardized spectral distributions such that any real colour can be matched by some combination of these primaries. The full spectrum now follows the curved line in the dia-gram, and a triangle of colours matched by three actual lamps lies inside this. A mathematical transformation is used to plot test results onto the graph.

The y axis indicates greenness, on a scale 0–1; the x axis indicates redness; and, because the sum of the three primaries is unity, the value of 1!x!y indicates blueness. Neutral colours lie close to the centre of the triangle.

Any colour of light, or a surface under a particular illuminant, or a TV screen phosphor, can be represented graphically by a point on the diagram, or numerically by the x and y coordinates. Mixtures can be predicted: the colours of all combinations of two coloured lights lie on a straight line between the graph points of the lights.

Several derivations of the original 1931 CIE diagram have been advanced. These in-clude the u, v and u!, v! forms, which aim to improve the perceptual uniformity so that a distance on the diagram represents a similar perceptual difference in every colour.

(31)

(32)

3 Seeing light and colour

The eye is not a camera. Although optical images are focused onto a light-sensitive surface at the back of the eye, these images are not what we perceive. In a series of transformations carried out first in the retina itself (the light-sensitive cells) and then in stages through to the visual cortex of the brain, the information is changed: the balance of brightness and colour is altered; attention is concentrated on small zones while large areas are unnoticed; images of the present scene are replaced by earlier images from memory. What we ‘see’ depends on experience and on what we have learnt, as well as on the physical structure of eye and brain.

But the physical system itself has many characteristics relevant to lighting design. They determine such factors as the range of brightness that can be perceived with comfort, sensitivity to visual changes, and the way coloured surfaces are recognized. This chapter introduces some concepts of vision that are used later in the book, especially in Chapter 9 on task lighting.

ADAPTATION

The eye alters in sensitivity in response to the light falling on it. This change can be seen when the pupil gets smaller in bright light, although contraction of the iris is not the main mechanism of adaptation but a fine adjustment for greater depth of field. The pupil changes in area over a range of about 16 to 1, but the eye is sensitive over a range of several million to one. It is the photoreceptors themselves, the light-sensitive cells of the retina, that make the adaptation. They contain pigments that are broken down by photons, releasing electrical energy and becoming less sensitive in the process. Once the light is removed, the pigments gradually regenerate so that sensitivity is regained. It is a process of self-regulation: the retina adapts itself to optimum sensitivity for the ambient lighting.

(33)

The bleaching out process is fast—a few seconds—while complete regeneration can take as long as an hour. Switching on the light when waking up at night gives only a short time of blinding glare; but it is many minutes before details of a dark room, such as a cinema, can be perceived after entering it from daylight. Figure 3.1 shows the time scale of dark adaptation. The step in the curve occurs because there are two types of photoreceptor in the retina, known as rods and cones. The rods are sensitive at low levels of light but are bleached out at the daytime levels, and can take 30 minutes to become fully dark-adapted; also, they do not give recognition of colour. The cones are sensitive only at brighter levels; three types of cone in the retina respond to different wavelengths and are the basis of colour vision.

Figure 3.2 Range of luminance discrimination.

Figure 3.2 illustrates how at every level of adaptation there is a limited range of brightness discrimination. Above this range, areas of very high luminance are glaring; within the central range the eye can sensitively distinguish between surfaces that differ only slightly in luminance; but below this all areas seem black, and little discrimination is possible. At the adaptation levels corresponding with normal room lighting the range of discrimination from the upper to the lower limit is about 1000:1. Brightness adaptation affects our ability to see small detail and small differences in contrast, and thus how we perform visual tasks. This is described in Chapter 9.

There is also visual adaptation to colour, especially the colour of ambient lighting: a surface appears to maintain the same hue when seen in light of gradually changing colour. We rarely recognize a change in room colours when daylight slowly varies, even though the colour of skylight can change greatly during the course of a morning. More complex types of visual adaptation, such as to tilt and to motion, also occur in everyday activities. For example, immediately after travelling in a car at high speed, movement at moderate speeds can seem very slow.

(34)

CENTRAL AND PERIPHERAL VISION

The alert eye is continuously in motion, and adjusts itself so that any object of interest is held within a central field of view about 2° across. This is about the height of a line of print in a book held a normal reading distance away. The ability to discern small detail —visual acuity—is greatest in this zone, and reduces by 50% only 5° from the centre. The small area of the retina in the central focus is called the fovea, and the central visual field is foveal vision.

Peripheral vision is especially sensitive to motion. Probably a relic of survival needs, there is automatic awareness of objects coming into the field of vision, a continuing monitoring of the ground when walking. Sudden changes of brightness, such as flashing lights, are noticed more when they occur at the edge of the visual field.

Foveal and peripheral vision are different and complementary. Having central vision alone is like being in a dark and unfamiliar room with a narrow-beamed torch: it is possible to scan about and pick up fine detail, yet be uncertain of the surroundings. To have peripheral vision alone is like entering a strange room that is dimly lit: the whole space can be perceived, but not in detail.

Colour and brightness sensitivities vary across the retina. The fovea comprises cone cells alone, and is therefore insensitive at low levels of light. Rods dominate the peripheral field, so colour is perceived poorly there. Movement of the eyes normally masks this partial insensitivity to colour, but if the gaze is fixed ahead and a coloured card is brought gradually into the field of view, the observer is aware of the moving object before the colour can be recognized.

WAVELENGTH, BRIGHTNESS AND COLOUR

When sunlight is split into a spectrum—as in a rainbow—the colours do not appear equally strong. The central bands of yellow and green have a much higher apparent brightness than the orange-red at one extreme or the purple-blue at the other. This is due primarily to the eye’s varying sensitivity, and it reflects the definition of light as human perception of radiation.

Figure 3.3 Sensitivity to spectral colour of the light-adapted eye: the

(35)

Figure 3.3 shows how the eye gives its maximum response to colours in the centre of the spectrum. Ultraviolet radiation, beyond the left of the curve, is not perceived by the eye: it is not ‘light’. A beam of radiation of constant power but gradually increasing wavelength would first look dull purple, and then a stronger blue, a bright green, a fading and reddening orange and dull crimson before finally becoming invisible again as the beam reached infrared.

The graph is the curve, the average visual response of the human eye in a light-adapted state. A similar function describes the average response of the human eye at low illumination, but this curve is shifted slightly to the left. Blue light, to the dark-adapted eye, is relatively brighter than red, but at the lowest brightnesses neither is perceived as a hue.

The relationship between the eye’s sensitivity and the spectral distribution of radiation, quantified by the curve, determines the luminous efficacy of a source. If all the electric power entering a lamp could be converted to radiation at the eye’s peak sensitivity, the luminous efficacy would be 683 lm/W. A lamp that has an output totally in the yellow-green region gives more light for the same radiated power than one giving a full spec-trum. There is therefore opposition between the luminous efficacy of a light source and its colour-rendering capabilities: to achieve a high efficacy not only should a lamp be designed to minimize power dissipated as conducted or radiated heat, or as radiation in the ultravio-let, but its visible output should be concentrated in the yellow-green. But for good colour rendering the lamp should have a balanced and continuous spectral output, which includes the colour extremes where the eye responds weakly. Chapter 4 describes this in more detail.

MIXTURES AND CONSTANCY

When several notes are played together on the piano, a chord is heard, not a single intermediate tone. With musical training the component notes can be identified and reproduced; the intervals between them even have names such as octave, fifth and minor third.

The ability to hear the component notes of a chord is not reflected in sight. Human vision is unable to separate the constituent wavelengths of a light source spectrum. For example, a mixture of red and green light falling on a white surface can be made to look the same yellow as the light of a low-pressure sodium lamp, because the eye cannot perceive the difference in spectral composition. This is of practical importance: paint colours can be mixed, and colour television and colour printing can occur, because of this absence of a one-to-one link between spectral wavelength and perceived hue.

But a different sort of discrimination can be made. Although a grey-painted surface and a white surface with lower illuminance may have the same luminance, they normally look different. Provided the light source is not concealed, the eye can separate the effects of surface and illuminant. The terms lightness (a term associated with surface reflectance) and brightness (which relates to luminance) are often used to distinguish this difference.

(36)

The word constancy is used to describe recognition of an unchanging characteristic when there is objective ambiguity. Size constancy occurs when an object is perceived as moving into the distance rather than shrinking, even though in both cases the same retinal image of the object is produced. Constancy of surface colour occurs when the eye can distinguish a piece of coloured material from white material illuminated by a coloured lamp. For this to happen the brain must be able to infer how the colour is produced.

The phenomenon of constancy indicates the sophistication of the eye and brain in their analysis of images. Clues from the whole visual field and from past experience are used subconsciously to select a visual pattern’s most likely interpretation. Perceptual constancy can, though, be overridden by unusual or deliberately illusory scenes, and when this happens the visual field is, literally, confusing.

SOME NON-VISUAL EFFECTS OF LIGHT

The 24-hour pattern of daylight and darkness triggers the human body’s cycle of alertness and sleep. An innate circadian rhythm is synchronized with daytime and night-time by a response to high levels of light (typically a few thousand lux) in early morning and late afternoon. The absence of these stimuli can cause poor nighttime sleep and difficulties of concentration by day; a high proportion of profoundly blind people report such symptoms. Exposure to periods of bright light can also aid the adjustment of shift workers to new hours, and can ameliorate jet-lag; and, in a related effect, high lighting levels can increase the alertness of workers, both on night shifts and in the afternoon when the body’s arousal is beginning its diurnal decline.

An extreme condition of light deprivation is seasonal affective disorder (SAD), a psychiatric depression that can occur in winter among those living in northern cities. The symptoms can be relieved by regular daily exposure to bright light.

(37)

(38)

4 Light from electricity

The first electric light, a carbon arc, was demonstrated in the middle of the nineteenth century. But it was the development of the incandescent lamp at about the same time by Joseph Swan in Great Britain and by Thomas Edison in the USA that heralded the start of electric lighting as we know it. Electric lamps have improved dramatically in quality, efficiency and convenience since the early times, but new light sources are still being developed.

Primarily there are two types of lamp used in buildings: incandescent and discharge. Each is available in a wide range, with variation in size, power, colour appearance, colour rendering, efficacy and operating characteristics. The lighting designer needs to be familiar with the different lamps available, and this information is best obtained from manufacturers’ catalogues.

INCANDESCENT LAMPS—THE HEATED FILAMENT

The incandescent lamp depends on passing an electric current through a wire to such an extent that it glows white hot. Tungsten wire is now used, but early lamps used carbon filaments; it is necessary to have a material with a high melting point so that it emits light for a reasonable length of time without breaking. The tungsten wire is usually coiled and coiled again to produce a fine filament. This is supported on two lead wires, which connect it to the electricity supply.

The bulb usually contains an inert gas to stop the filament oxidizing; sometimes a vacuum is used. But gradually, as the lamp operates, tungsten evaporates from the filament and is deposited on the inside of the bulb. The filament becomes thinner and eventually breaks. An inert gas filling retards this process and ensures a reasonable lamp life, typically about 1000 hours. The actual life varies within a batch of lamps, and although a mean value can be quoted this may not be helpful. Lamp manufacturers often give the life as the point when they expect a particular percentage of lamps to have failed.

(39)

Figure 4.1 Variation with voltage of the life and light output of an

incandescent lamp.

The performance of the lamp is determined primarily by the operating temperature of the filament. As this increases, there are three important consequences: the colour of the light produced becomes whiter; there is an increase in luminous efficacy (the amount of light produced in relation to the energy input); and the life of the lamp is shortened. A small change in the electricity supply voltage therefore has a major effect. This is shown in Fig-ure 4.1. An increase in applied voltage of 5% (for instance, from 240 to 252 V) causes an increase in luminous output of 20%, but the operating life is halved. Conversely, underrunning an incandescent lamp increases life significantly but reduces the light output.

The colour appearance of an incandescent lamp is described by its colour

temperature, measured in kelvins. This is related to the temperature of the filament. A

normal tungsten filament lamp has a colour temperature of about 2800 K—a relatively warm colour appearance. The spectrum of radiant energy from the lamp has the shape shown in Figure 4.2. Its smooth continuity means that its colour rendering is good: the light allows fairly accurate discrimination between one surface colour and another. But the output is not uniform across the visible range: there is far more energy in the red than the blue wavelengths. This causes the warm appearance but means also that red surfaces illuminated by the lamp may look brighter than blue surfaces of the same reflectance.

(40)

Figure 4.2 Spectral output of an incandescent lamp.

Most of the radiant output—over 90% of it—is not in the visible part of the spectrum; the peak occurs far into the invisible infrared region. The majority of energy from an incandescent lamp is emitted as radiant heat, which is why its luminous efficacy is low, typically 12lm/W.

A major development of the incandescent lamp followed the discovery of the halogen cycle. In this process the tungsten that evaporates is redeposited back onto the filament. It occurs when a halogen, usually iodine or bromine, is included with the filling, and the gas temperature is controlled. The performance effect of this is an increase of lamp life to typically 2000 hours, an increase in colour appearance to 3000 K, and an increase in efficacy to typically 20 lm/W.

Incandescent lamps are produced in many different bulb shapes and finishes: clear and diffuse, as well as a range of colours. The bulbs can be partly silvered and formed into an integral reflector to make spotlamps. A range of these is available, with various beam shapes and intensities. Small compact tungsten halogen lamps, particularly those operating at low voltage (usually 12 or 24 V), can be used with precise optics to provide accurate light control. A family of very small low-voltage spotlamps is available. Used extensively for display lighting, these lamps are operated on mains voltage with small transformers.

DISCHARGE LAMPS—THE GLOWING GAS

Light can be produced by an electric discharge in a gas-filled transparent tube. The discharge is started by applying a high voltage across the electrodes at each end. This ionizes the gas filling, enabling an increasing current to flow, and resulting in further ionization. The radiation produced depends on the materials in the tube and on the gas pressure. Its spectrum is dis-continuous, and comprises bands of radiation at specific wavelengths. Phosphor coatings on the inside wall of the tube may be used to absorb some of the radiation and re-emit it at different wavelengths—especially to convert ultraviolet radiation to energy in the visible range.

(41)

With all discharge lamps additional equipment is required in the electrical circuit. This produces an initial high voltage to start the discharge, then limits the current during operation and controls the power factor. The power factor depends on the relationship between voltage and current in an a.c. circuit, and affects the efficiency of the equipment. The combined efficacy of a lamp and its control circuit determine the energy efficiency.

The colour appearance of a discharge lamp is specified by the correlated colour

temperature (CCT): the temperature in kelvins of the black-body radiation that appears

closest to the colour appearance of light from the lamp. A colour temperature below 3300 K is often describes as warm, between 3300 and 5300 K as intermediate, and over 5300 K as cold.

The CIE general colour rendering index (R

a ) provides a measure of a lamp’s colour

rendering quality around the hue circle, quantified on a scale from 0 to 100. Because it has a continuous spectrum the incandescent lamp is used as a reference, and is assigned R

a=100. Lamps with an R a greater than 90 are considered to be very good, and are used

where accurate colour matching or discrimination is required. Those with an R

a in the

range 80–89 are appropriate where accurate colour judgement is necessary or where good colour judgement is required for reasons of appearance. Lamps with an R _a below 80 should be used only where colour quality is of little importance.

The most commonly used discharge lamp is the low-pressure fluorescent tube. This uses primarily a mercury discharge, which emits a large part of its energy as ultraviolet radiation. The inside wall of the lamp tube is lined with a phosphor powder, which absorbs the ultraviolet and re-radiates the energy in the visible spectrum.

Since its introduction in the early 1940s the lamp has gone through many developments. Much work has been done on phosphor compositions, and now it is possible to have lamps with colour appearances ranging from warm to cold. The lamp is efficient, with a high efficacy, although the actual value depends on the phosphor composition. Typically, a modern fluorescent lamp using multi-phosphor technology with good colour rendering (R _a=80) has an efficacy of 80–100 lm/W. The life is usually 8000–10 000 hours, but the light output reduces gradually with age.

Originally the electrical circuit for fluorescent lamps included a wire-wound inductive ballast, a starter switch and a power factor capacitor. Recently, electronic circuits have been introduced. These operate lamps at very high frequencies, 20–30 kHz rather than the 50 Hz of the normal mains supply. The very high frequency improves energy efficiency, but comfort is also improved because lamp flicker is undetectable. A further advantage of electronic circuits is the ability to regulate or to dim the light output. The light from the lamp can be controlled to vary with daylight or dimmed for special conditions.

The fluorescent tube has near instant switch-on and restrike, especially with electronic control. The light output increases from switch-on, usually reaching its maximum after a few minutes, but the change in output after switch-on will be hardly noticed.

The output of a fluorescent lamp depends on the temperature of the coolest spot on the bulb wall and therefore on the ambient temperature. If a fluorescent lamp is to be used at low ambient temperatures (in a cold store, for example) or at high ambient temperatures (such as in a bakery) then the lamp light output will be different in the two extremes from normal operating temperatures.

(42)

In the compact fluorescent lamp (or CFL) the lamp tube is folded and combined with an integral control circuit to form a discharge lamp with a volume similar to that occupied by an equivalent incandescent lamp. The early versions, developed during the 1980s, were heavy because of the integral control gear, but they formed an important breakthrough in lamp development. There has since been considerable improvement, and now it is possible to have CFLs with either integral or separate control gear. It is likely that this type will become one of the most commonly used lamps, and a direct replacement of incandescent lamps for many purposes.

Compact fluorescent lamps have good colour-rendering properties and come in a range of different colour appearances. Lamp life is high, typically 8000–10 000 hours, but the output decreases with age. Efficacy is typically 50–70 lm/W: this is high by comparison with incandescent lamps but less than that of the tubular fluorescent.

Cold-cathode lamps have an unheated filament but require high-voltage control gear.

They have long, thin tubes, which can be bent into signs or shaped to fit architectural features. Their life is long, 30 000 hours, but their efficacy is typically about 50 lm/W. When filled with gases other than mercury vapour, and the phosphor coating omitted, they can produce light in various colours.

The high-pressure discharge lamp has a small discharge tube contained within a tubular or elliptical outer bulb; there is not necessarily a fluorescent coating. It is much smaller than the tubular fluorescent lamp but operates similarly; it also requires ancillary electrical control equipment to initiate the discharge, control the current, and correct the power factor. A high-pressure discharge lamp takes a few minutes to achieve full light output; if switched off and back on again, there is a delay before the lamp cools sufficiently for the arc to form again, unless hot restrike control gear is provided.

The high-pressure mercury lamp was the first to be introduced, in the 1930s. It was used almost exclusively for street lighting, because although it was efficient compared with incandescent sources it had a very poor colour appearance and rendering. The lamp comprises a quartz arc tube contained within either an elliptical or a reflector-shaped outer bulb. It has long life, typically 10 000 hours, and a reasonable efficacy, 40–60 lm/ W. In the lamp’s basic form the colour performance is poor, but this is significantly improved in versions of the lamp with phosphor coating inside the bulb.

In the 1960s a development of the mercury high-pressure discharge lamp occurred that raised its colour quality to the point where it could be used where colour rendering and appearance were important. In the metal halide lamp, halogens are added to the mercury vapour. To do this successfully requires advanced manufacturing technology. Initially the colour appearance of some lamps tended to change through life: this was particularly noticeable when several were used in the same installation and could be easily compared, but the technolo-gy has been developed intensively and the major lamp makers have minimized this effect. The lamp comes in a range of shapes and sizes, including versions with integral reflectors. Some lamps have very small arc tubes, which makes them ideal for precise optical control lumi-naires, such as spotlights for sports lighting. The lamp life is typically 8000–10 000 hours with an efficacy of 70–100 lm/W. In addition to the ‘white’ metal halide lamp, coloured lamps, particularly blue and green, are produced by some manufacturers for decorative purposes.

The other main group of lamps is based on a discharge through sodium vapour. The

low-pressure sodium lamp has a bright orange monochromatic light. It is used only

(43)

application. The low-pressure sodium lamp is available in a range of sizes, all with tubular bulb shapes. Typically it has a life of 8000–10 000 hours and an efficacy of 100–200 lm/W.

Lamp scientists had known for some time that if a sodium discharge at high pressure could be created, a much improved colour performance could be achieved. As the pressure of a sodium discharge is raised, the spectrum of monochromatic radiation at low pressure expands to produce a broadband distribution. The problem was to find a light-transmitting material that could contain the highly corrosive sodium at high pressure. In the 1960s a translucent ceramic material, sintered alumina, was developed. Research has continued on this, making possible further increases in arc pressure and hence even better colour performance. Currently the high-pressure sodium lamp is available in a range of sizes and bulb shapes. The best colour versions are described as ‘White SON’. The lamp has a long life, typically 8000–10 000 hours, with an efficacy of 70–120 lm/W.

The development of lamps is a continuing process, and two new types of lamp have been recently introduced. The mercury induction lamp depends on energizing a mercury discharge using a magnetic field. Because this eliminates the need for electrodes, which deteriorate with time, the lamp life can be extremely long: typically 60 000 hours is quoted by the manufacturers. The lamp is basically a fluorescent tube and has a similar colour quality, with an efficacy of approximately 60 lm/W.

The sulphur microwave lamp is an electrodeless source, which uses microwaves to create light from a sulphur and argon bulb filling. The prototypes produce 450 000 lm for an energy consumption of 5.9 kW: an efficacy of approximately 76 lm/W. The life is estimated to be 10 000 hours, based on the life of the magnetron that generates the microwaves. The spectral distribution is continuous across the whole spectrum range, with reasonable colour performance. With a source of this power a system is necessary to distribute the light within buildings; the sulphur lamp cannot be used as direct replacement for existing sources until small sizes can be manufactured. A 1 kW version is commercially available.

WHICH LAMP?

Table 4.1 lists the characteristics of some incandescent and discharge lamps. Each type of lamp is produced in many versions, varying in power, efficacy and colour, so it is necessary to use manufacturers’ data when designing. The values in the table given for efficacy are those for a lamp plus its control equipment (if any); the design life is a relative value, taking into account the probability of failure and long-term reduction in light output.

Table 4.1 Typical lamp characteristics

There is a wide range of lamps available in each group, varying in characteristics. Manufacturer’s data must be used when specifying particular lamps.

Standard incandescent

Circuit luminous efficacy 12 lm/W

(44)

Colour R

a 100, 2800 K

Control gear required? No

Tungsten halogen

Circuit luminous efficacy 20lm/W

Design life 2000 hours

Colour R _a 100, 3000 K

Control gear required? Transformer (for low voltages)

Tubular fluorescent

Circuit luminous efficacy 80lm/W

Colour R _a 85, 2700–6500 K

Control gear required? Yes

Compact fluorescent

Colour R _a 80, 2700–4000 K

High-pressure mercury fluorescent

Design life 10 000 hours

Colour R

a 60, 3300 K

Metal halide

Colour R _a 80, 4000 K

High-pressure sodium

Design life 12 000 hours

Colour R

a 60, 2400 K

Choosing a lamp for a particular application means that a balance has to be made between different requirements. Table 4.2 lists some of the points to be considered.

Table 4.2 Short checklist for choosing a lamp

Power and output

Power rating (watts), light output (lumens) and efficacy (lumens/watt)

Dimensions and durability

Bulb size and shape, burning position, mechanical strength

Colour

Colour rendering index (R

(45)

Lifetime

Lamp life (hours) and lumen depreciation

Control

Type of control gear required, lamp run-up time (minutes), restrike performance and ability to dim

Thermal

Lamp envelope temperature, lamp operation with respect to ambient temperature

CONTROLLING THE LIGHT

A luminaire has several functions. It must give the lamp physical support and protection; it must enclose electrical connections; it must support any associated control gear; it must have an appearance that is architecturally appropriate. Above all it has to provide optical control so that the light output has the required distribution.

Several techniques for redirecting light are available to the luminaire designer, and include the following:

• Obstruction. This is the use of masks to control the light. An example is the provision of louvres to limit discomfort glare caused by a view of bare lamps. Another example is the simple drum fitting, which obstructs light to the sides but allows light to travel both up and down. Elements of a building (such as coves, cornices and pelmets) can be used to provide obstruction, integrating the luminaire with the building itself.

• Reflection. A flat mirror redirects a ray of light so that the angles of incidence and

reflection are the same, relative to the normal of the mirror. If the reflector is curved, the rays from the source can be focused. A non-specular surface scatters light; if the surface is perfectly diffusing, or ‘matt’, it scatters light evenly in all directions, so the luminance of the surface is constant from all directions of view. Together, the shape of the surface, its specularity and its reflectance can be used to determine the luminaire’s beam shape. Spotlights used in buildings depend primarily on a focused reflection.

• Refraction. The direction of a beam of light is changed at the junction of two

transparent materials, such as air and glass. When a ray at an oblique angle passes from air into glass, its direction bends away from the surface of the denser medium; it will change again when it emerges from the other surface of the glass, unless it strikes the junction at right angles. Lenses are refractors that focus light into a particular direction. Transmitting materials range from clear to diffusing; the degree of scatter, the shape of the surfaces and the transmittance of the material affect the final beam shape and intensity. The control of light by stage spotlights and street lighting lanterns is accomplished primarily by refraction.

Most luminaires use a combination of different forms of optical control to produce the light distribution required for a particular application.

(46)

Figure 4.3 Luminous intensity distribution from a fluorescent tube

luminaire.

The photometric performance of a luminaire is frequently given by two measures: the light

output ratio (LOR), and the luminous intensity distribution, which is often called the polar curve. The light output ratio is the proportion of the lamp light output that emerges

from the luminaire; it is expressed as a fraction or a percentage. For an installation to have a high energy efficiency the light output ratio of its luminaires should be high.

The intensity distribution describes the pattern of light emerging from the luminaire. It is frequently presented as a graph such as that in Figure 4.3, which illustrates the performance of a fluorescent fitting. The distance of the curve from the centre is proportional to the intensity in that direction. The two parts of the graph represent the intensity along the axial and transverse axes. The intensity distribution is usually described in terms of a lamp light output of 1000 lm (therefore intensity is given in units of candela/1000 lm). The reason for this is that the luminaire may be available for differ-ent lamp sizes; the user scales the intensity values according to the lamps used.

The photometric information provided by the luminaire manufacturer refers to a new and clean luminaire. Light output reduces through life with the changing output of the lamps and accumulation of dirt.

(47)

Luminaires are produced by the lighting industry for a very wide range of applications. The designer needs to balance several factors when choosing luminaires for a particular purpose, and to anticipate the lifetime use of the installation. Table 4.3 lists some points to be considered.

Table 4.3 Short checklist for choosing a luminaire

Lighting performance

Light output ratio and intensity distribution

Electrical control requirements

Switching, dimming, automatic control linked to daylight or occupancy

Safety

Mechanical, electrical and thermal safety during installation and maintenance as well as during normal use

Installation requirements

Support, fixing, power supply, access

Maintenance

Frequency of cleaning and lamp replacement, eventual replacement of the complete installation

Appearance

The Design of Lighting

The design of lighting

Peter Tregenza and David Loe

E & FN SPON

An Imprint of Routledge

Contents

List of tables

Preface

The Technical Background

Describing light

2

Describing colour

3

Seeing light and colour

4

Light from electricity